1

JOYCE J. EVANSet al.

GENOMIC DIVERSITY OF STREPTOCCOCUS AGALACTIAE ISOLATES FROM MULTIPLE HOSTS AND THEIR INFECTIVITY IN NILE TILAPIA

JOYCE J. EVANS1, PHILLIP H. KLESIUS2, JOHN F. BOHNSACK3, DAVID J. PASNIK1, JULIO C. GARCIA2 ,APRIL A. WHITING3 AND

CRAIG A. SHOEMAKER2

  1. Aquatic Animal Health Research Laboratory, United States Department of Agriculture, Agricultural Research Service, 118B Lynchburg Street, Chestertown, MD, USA 21620

Email:

  1. Aquatic Animal Health Research Laboratory, United States Department of Agriculture, Agricultural Research Service, 990 Wire Rd, Auburn, Alabama, USA 36831
  2. Department of Pediatrics, University of UtahHealthSciencesCenter, Salt Lake City, Utah, USA

Abstract

Our laboratory has conducted multiple studies to investigate the genomic diversity of GBS isolates from different phylogenetic hosts and geographical regions. We have examined fish and dolphin GBS strains using phenotypic, serological typing and multilocus sequence typing (MLST) techniques and compared these to bovine and human GBS isolates. Studies were also conducted on infectivity of fish, dolphin, bovine and human GBS isolates in Nile tilapia, Oreochromis niloticus. A previously unreported fish capsular serotype, Ia, was discovered for tilapia isolates originating from Brazil, Israel and the U.S.A, and mullet, seabream and dolphin isolates from Kuwait. Serotype Ib was noted for other tilapia isolates from Brazil and Honduras. Sequence typing of isolates produced six sequence types (ST-7, ST-257, ST-258, ST-259, ST 260 and ST-261), the latter five ST’s representing allelic designations and allelic combinations unique to the S. agalactiae MLST database. Kuwait isolates (Ia, ST-7), although largely unrelated to the majority of bovine and human GBS strains, appear to share a common ancestry. Kuwait GBS isolates shared the same allelic profile, sequence type and capsular serotype as that reported from human GBS strains from Japan (Ia, ST-7). Genomic diversity existed between Kuwait GBS isolates and those from other geographical areas. Tilapia GBS isolates from Brazil, Israel, Honduras and the U.S.A. are part of a clonal complex and are unrelated to bovine and human GBS, thus representing a distinct genetic population. In experimental studies using fish isolates from different geographical regions, a Brazil GBS tilapia isolate was significantly more pathogenic to Nile tilapia than non tilapia isolates from Israel and Kuwait. However, all fish isolates caused mortalities at doses between 101 and 106 CFU/fish. Bovine serotype Ia, II and NT GBS isolates of unknown MLST type were not found to be infective to Nile tilapia although a human GBS isolate caused mortality in Nile tilapia. These experimental infectivity studies indicate enhanced virulence of fish GBS isolates to fish regardless of geographical origin. Here we summarize findings from GBS genomic and infectivity studies from our laboratories.

INTRODUCTION

Treptococcus agalactiae, the Lancefield group B Streptococcus (GBS), has a broad host range and can be pathogenic to numerous animals, including fish. GBS is most recognized for causing cattle mastitis and human neonatal meningitis, although it also causes fatal meningo-encephalitis in fish. Natural tilapia infections resulting from GBS are increasing in occurrence. Despite the scant literature on GBS hosts and distribution, our laboratory has characterized GBS from various North, Latin and Central American and Asian tilapia farms. Three species of tilapia have been naturally infected with group B Streptococcus agalactiae in Israel, Thailand, and Brazil. Natural S. agalactiae infections in tilapia cultured in the U.S.A have not been published. Despite the initial reports of non-hemolytic GBS causing fish kills in cultured freshwater golden shiners in the U.S.A and later rainbow trout Oncorhynchus mykiss in Israel, the majority of natural infections of S. agalactiae in the U.S.A (Robinson and Meyer 1966; Plumb et al. 1974), Japan (Eldar et al. 1994), and Kuwait (Evans et al. 2002; Duremdez et al. 2004) have involved non-tilapine estuarine and marine fish. The reader is referred to Evans et al. (2006a; 2006c) for reviews of S. agalactiae in tilapia and non-tilapine species.

GBS worldwide incidence and epidemiological and zoonotic considerations necessitate an undertaking of a comparison of GBS isolates from different phylogenetic hosts. The application of molecular and genetic techniques to human and bovine GBS have resulted in better discrimination between their genetic relatedness geo-spatially (Jones et al. 2003; Bohnsack et al. 2008). Genetic variations in fish GBS isolates and differences between fish, human and bovine GBS isolates have only recently been explored. Piscine and dolphin GBS isolates have recently been reported to share the same serotype (Ia) and sequence type (ST-7) as that reported from human GBS adult and neonatal invasive strains from Japan and North America (Evans et al. 2008a in press). This study (Evans et al. 2008a in press) examined geo-spatial relationships among GBS fish isolates using phenotypic, serological and multilocus sequence typing (MLST) molecular techniques and compared these to previously-characterized human and bovine GBS sequence types (STs) (Bohnsack et al. 2008) to help elucidate their phylogenetic similarity and differences. Here we summarize results from GBS genomic and infectivity studies from our laboratories.

MATERIALS AND METHODS

Identification and characterization of GBS strains

The genomic study employed 30geographically diverse isolates from fish, bottlenose dolphin, bovine and human sources (Table 1) (Evans et al. 2008a in press). Isolates were grown on 5% sheep blood agar (SBA) (Remel, Lexena, KS). Conventional biochemical tests (hemolysis, CAMP, pyrrolidonylarylamidase (PYR) reaction, and hippurate, urea, starch, arginine and esculin hydrolysis) were performed as described by MacFaddin (2000). The Voges-Proskauer (VP) reaction and fermentation of sorbitol, trehalose, ribose, inulin, mannose, xylulose and lactose were derived from the API 20 multi-test system following manufacturers' instructions. Lancefield grouping was performed with type B antisera (Oxoid, Basingstoke, United Kingdom).

Capsular serotyping was performed twice on all strains in the laboratories of John Bohnsack and repeated in a second laboratory by Phillip Klesius (USDA, ARS Aquatic Animal Health Laboratory, Auburn, AL36830USA). Capsular serotype was determined by group B streptococcal typing antisera, directed against capsular type Ia, Ib, II, III, IV, and V from Denka Seiken, Tokyo, Japan. Serotypes were additionally identified by a multiplex PCR assay performed twice on some isolates. The protocol was modified from Poyart et al.(2007) and determined serotypes Ia, Ib, or II according to capsular polysaccharide gene (cps) amplicon base pair size.

Table 1. Collection of GBS strains

Isolates (No.) / Country / Host / Serotype / Sequence type
01-ARS-KU-MU-2 NA, 3 BR, 9 BR, 15 NA, 24 BL, 24 BR, 34 E,
KU-SB-37 HK, 38 BR,1 NA, 9 NA, 11 NA, 11 BR, 19 HK (14) / Kuwait / Mullet 1 / Ia / 7
01-ARS-KU-SB-1HK (1) / Kuwait / Seabream / Ia / 7
03-ARS-BZ-TN-05 (1) / Brazil / Nile tilapia / Ib / 257
03-ARS-BZ-TN-06 (1) / Brazil / Nile tilapia / Ia / 257
LADL05-108A (1) / Honduras / Nile tilapia / Ib / 260
LADL00-351A (1) / USA / Nile tilapia / Ia / 259
IS-ET-09-03 (1) / Israel / Hybrid striped bass / Ib / 258
ATCC 51487 (1) / Israel / Hybrid tilapia / Ia / 261
01-ARS-KU-BD-MU (1) / Kuwait / Dolphin 2 / Ia / 7
841254, 510036, 560249,
630635, 510012, 510029 (6) / Japan / Human 3 / Ia / 7
ATCC 13813 (1) / United Kingdom / Bovine 4 / II / 61
ATCC 27956 (1) / USA / Bovine / NT / 23

Fish isolates (n=21) originated from infections in Klunzingeri mullet, Liza klunzingeri and seabream, Sparus auratus (Kuwait) (Evans et al. 2002), hybrid striped bass, Morone chrysops x Morone saxatilis (Israel), Nile tilapia, Oreochromis niloticus (U.S.A, Brazil, Honduras), and hybrid tilapia, Oreochromis spp. (Israel) (Eldar et al. 1994). Nile tilapia GBS isolates were provided by Dr. John Hawke (Louisiana Veterinary Diagnostic Laboratory, LouisianaStateUniversity, Baton Rouge, Louisiana, U.S.A. and Dr. Rogério Salvador (Parana State, Brazil). Dr. Dina Zilberg (The Albert Katz Department of Dryland Biotechnologies, Ben-GurionUniversity of the Negev, Israel) provided the hybrid striped bass isolate.

2 The bottlenose dolphin isolate (Kuwait) was obtained as described in Evans et al.(2006b).

3 Six ST-7, serotype Ia GBS isolated from humans in Japan were provided by Dr. Shinji Takahashi (Division of Microbiology, Joshi-EiyohUniversity, Chiyoda, Sakado, Saitama 350-0288, Japan).

4 References strains were obtained from the American Type CultureCollection (ATCC),Washington, DC, USA. (ATCC 13813 and 27956 GBS from bovine species and ATCC 51487 deposited as S. difficilis, but later identified as S. agalactiae, from diseased tilapia).

Multilocus sequence typing (MLST)

MLST, allelic assignment and sequence typing was performed as previously described (Jones et al. 2003). Allele sequences can be found at . MEGA softwar (Kumar et al. 2004) was used to construct an unrooted dendrogram showing the relationships among fish, dolphin, bovine and human STs as compared to an extensive collection of human (n = 899) and bovine isolates (n = 63) from North America (Bohnsack et al. 2008). The eBurst program (Feil et al., 2004) was used to look for evidence of clonal complexes among these isolates (Evans et al.2008a in press).

Infectivity of GBS isolates

Experimental infectivity trials were performed on Nile tilapia in USDA laboratories as using Kuwait mullet and seabream GBS isolates (Evans et al. 2002), an Israeli hybrid striped bass isolate and a Brazilian Nile tilapia isolate (ARS unpublished), Kuwait bottlenose dolphin isolate (Evans et al. 2006b), bovine isolates (Garcia et al. 2008 in press) and a Japanese human GBS isolate (2008b in press) and compared to other published reports (Table 2). Briefly, Nile tilapia (n > 500; weight ranging from 3 to 100 g) were injected intraperitoneally with doses ranging from 101 to 109 colony-forming units (cfu) S. agalactiae/fish and maintained at approximately 30˚C. Control fish were injected with TSB. Fish were monitored for up to 14 days post-challenge, and culture samples were aseptically obtained from the nares, brain, anterior kidney, and posterior intestine of dead fish to confirm the presence of S. agalactiae.

RESULTS AND DISCUSSION
Phenotypic characterization

Phenotypically, all isolates were Lancefield group B, catalase-negative, Gram-positive cocci, appearing in pairs and chains. Kuwait mullet, seabream and bottlenose dolphin isolates, human isolates and one bovine reference isolate (ATCC 27956) were -hemolytic. Tilapia isolates from Brazil, Honduras, U.S.A, Israel and one bovine reference isolate from the United Kingdom were non-hemolytic. All isolates were PYR negative. None of the isolates hydrolyzed urea or starch or fermented sorbitol, mannose, or xylose. All isolates hydrolyzed hippurate, fermented ribose and were positive by the VP reaction. Non-hemolytic tilapia and hybrid striped bass isolates phenotypically varied from -hemolytic fish and human isolates in that these isolates did not ferment trehalose and were CAMP negative. In contrast, the non-hemolytic bovine reference isolate was CAMP positive while the -hemolytic bovine isolate was CAMP negative although both fermented trehalose. The majority of isolates hydrolyzed arginine. Esculin hydrolysis was primarily negative with the exception of two Kuwait fish isolates and one bovine reference isolate (ATCC 27956). Likewise, the majority of isolates did not ferment inulin, with the exception of two Kuwait fish isolates, or lactose. Lactose fermentaion was noted in one Kuwait mullet and one Brazil tilapia isolate and in both bovine isolates.

Relatedness between ST and capsular serotype

The nine capsular serotypes of GBS (Ia, Ib, II-VII) have been the gold standard in differentiation between GBS isolates. Two capsular serotypes, Ia and Ib, were determined for the fish isolates (Table 1). Serotype Ia, a previously unrecognized serotype for fish GBS isolates, was the most common serotype occurring in all of the Kuwait fish isolates, dolphin isolate and Japan human isolates and in one tilapia species isolate each from Israel (ATCC 51487), Brazil (03-BZ-TN-05) and U.S.A (LADL00-351A). Three fish isolates were serotype Ib (one isolate each from Israel (IS-ET-09-03), Brazil (03-BZ-TN-06) and Honduras (LADL05-108A). Capsular serotype was not necessarily restricted to a specific ST. One ST (ST-257) contained isolates with different serotypes, Ia and Ib. Reference bovine isolates were serotype II (ATCC 13813-United Kingdom) and nontypable (ATCC 27956-U.S.A.). Serotype Ib and no serotype (NT or nontypable) have been previously reported for piscine GBS isolates. Wilkinson et al.(1973) serotyped non-hemolytic piscine GBS strains obtained from Alabama, U.S.A (1 isolate) and Arkansas, U.S.A (3 isolates) (Robinson and Meyer 1966) and invasive neonatal GBS (4 isolates) using agar gel diffusion techniques and found these to be type Ib. Eldar et al. (1994) found the Israeli S. difficile isolate to be non-typable. Vandamme et al. (1997), using a coagglutination assay (DARO group B Streptococcus serotyping test; DAKO A/S Glostrup, Denmark) later reported the capsular serotype of Eldars' S. difficile isolate (LMG 15799) as type Ib. Piscine non-hemolytic GBS isolates from Brazil (Salvador et al. 2005) and beta-hemolytic GBS isolates from Kuwait (Evans et al 2002; Duremdez et al. 2004) and Thailand (Suanyuk et al. 2005) were not previously serotyped. Evans et al. (2008a in press), in replicate capsular serotyping of the Israeli ATCC type isolate (51487) revealed the isolate to be Ia. These results are inconsistent with those reported by Eldar et al. (1994) and Vandamme et al.(1997). It appears this isolate may have been mistyped or differences in typing antiserum or technique employed play a role in disparate capsular serotype results.

Relatedness of GBS isolates and variations in the MLST

The 21 fish isolates were categorized into six sequence types (ST’s), four of which were only identified once (ST-258, ST-259, ST-260, ST-261) (Table 1) in fish isolates. All of the Kuwait fish isolates and dolphin isolate and human isolates were represented by ST-7. Six fish isolates were represented in four newly recognized ST's. Both Brazilian Nile tilapia isolates were assigned to ST-257. One Israeli hybrid striped bass isolate was represented by ST-258 while another Israeli tilapia isolate (ATCC 51487) was represented as ST-261. The U.S.A. and Honduras Nile tilapia isolates were assigned to ST-259 and ST-260, respectively. Bovine reference isolates were ST-23 and ST-61. A dendrogram shows genetic relatedness between human, bovine and fish isolate STs (Figure 1).

Figure 1. Dendrogram illustrating the phylogenetic relationship of the piscine isolates to previously characterized human and bovine GBS ST’s (Bohnsack et al. 2008). Clonal Complexes (CC) are outlined with brackets, as are the ST’s of piscine isolates.

MLST demonstrated the unrelatedness of the different geographical fish GBS isolates to each other or to human or bovine STs. Kuwait fish and dolphin isolates appear to be from a single clone and largely unrelated to the majority of human GBS strains. However, Kuwait GBS fish isolates shared the same allelic profile, sequence type and capsular serotype as that reported from one human GBS carried strain and two neonatal invasive strains from Japan (Ia, ST-7) (Jones et al. 2003). None of the other piscine isolates examined conformed to ST's or allelic profiles reported from humans. Five of the piscine isolates causing epizootics from 4 different geographical regions (Brazil, Israel, U.S.A and Honduras) were more diverse and had alleles and allelic sequences not previously deposited in the MLST database. These isolates required assignment of numbers to the alleles and assignment of an ST to the combination of alleles. Based on dendrogram results, these isolates appear to be part of a clonal complex. An examination of the GBS population structure demonstrate that at least two divergent populations of GBS are capable of causing epizootics in fish, and that one of these populations (ST-7) is capable of also causing human infections, although rarely and apparently in restricted geographical areas.

Infectivity of GBS isolates

Experimental infectivity studies indicate enhanced virulence of fish GBS isolates to fish regardless of geographical origin. Regardless of the fish isolate used, some fish in each experimental challenge study exhibited clinical signs such as: going off feed, lethargy, exophthalmia, C-shaped body posturing, erratic swimming, and fecal string formation. The Kuwait mullet and seabream GBS isolates caused 90-100% mortality in Nile tilapia and had a Lethal Dose (LD50) of 1.9 X 103 cfu/fish (Evans et al. 2002) (Table 2). In subsequent infectivity studies, a Brazil GBS tilapia isolate (03-ARS-BZ-TN-05) was significantly more pathogenic to Nile tilapia than non tilapia isolates from either Israel (03-ARS-IS-ET-09) and Kuwait. However, all fish isolates caused mortalities at doses between 101 and 106 CFU/fish. Figueiredo et al.(2007)also indicated increased virulence with Brazilian GBS tilapia isolates. In other studies, the Israeli S. difficile tilapia isolate (Eldar et al. 1995) increased in virulence after a series of in vivo passages decreasing the LD50 to 102 cfu/fish. Chang and Plumb ( 1996) noted increased susceptibility of Nile tilapia to GBS isolates from non tilapia fish species with elevated salinities and temperature suggesting environmental conditions are important to GBS susceptibility.

Bovine serotype Ia, II and NT GBS isolates of unknown MLST type were not found to be infective to Nile tilapia, although a serotype Ia ST-7 human GBS isolate (ID# 510012) obtained from a human neonatal meningitis clinical case in Japan, caused mortality in Nile tilapia (Evans et al 2008b in press). The human isolate was virulent in tilapia at 102, 103 and 107 cfu/fish and morbid fish were culture positive for GBS. In tilapia infectivity studies using a dolphin GBS isolate, disease signs and 90% mortality were noted within six days post-inoculation with 107 cfu/fish. These latter experimental findings suggest the possibility of mammalian GBS infectivity to fish and origination of GBS-induced fish epizootics from mammalian sources. While only one human and dolphin isolate was studied here, it is possible that other isolates with different ST’s or isolates repeatedly passed through fish may display increased virulence. The relationship between capsular serotype, ST and infectivity of GBS isolates to different hosts is an area deserving more research.

Table 2. Experimental infectivity of S. agalactiae from multiple hosts to Nile tilapia1

Isolate (s) / Host Origin / Country / Dose
cfu/fish / % Mortality / Reference
01-ARS-KU-MU-11BR / Mullet / Kuwait / 1 x 107 / 100 / Evans et al. 2002
01-ARS-KU-SB-37BR / Seabream / Kuwait / 1 x 107 / 90 / Evans et al. 2002
03-ARS-IS-ET-09 / Striped bass / Israel / 101-106 / 20-90 / USDA unpublished
03-ARS-BZ-TN-05 / Nile tilapia / Brazil / 101-106 / 20-100 / USDA unpublished
Lake / Speckled trout / USA / 2 x 106 / 27-100 / Chang and Plumb 1996
DL 805 / Gulf killifish / USA / 2 x 106 / 33-100 / Chang and Plumb 1996
MS91-452 / Channel catfish / USA / 2 x 106 / 3-63 / Chang and Plumb 1996
Not reported / Nile tilapia / Thailand / 101-108 / 20-90 / Suanyuk et al. 2005
ND 2-22 (ATCC 51487) / Hybrid tilapia / Israel / 107 / 50 / Eldar et al. 1995
ST 20-06 / Hybrid tilapia / Brazil / 6.4 x 101 / 50 / Figueiredo et al. 2007
MO-55100, 59918, 59998, 61185, 61321 / Bovine / USA / 109-1010 / 0 / Garcia et al. 2008 in press
ATCC 27956 / Bovine / USA / 105 , 107 / 0 / USDA unpublished
01-ARS-KU-BD-MU / Bottlenose dolphin / Kuwait / 1 x 107 / 90 / Evans et al. 2006c
510012 / Human / Japan / 102-107 / 11.7 / Evans et al 2008b in press

Fish were challenged by intraperitoneal injection with the specified isolate at he indicated doses and observed for mortalities for up to 14 days post-challenge.